RESUMO
Single-cell bisulfite sequencing (scBS) is a technique that enables the assessment of DNA methylation at single-base pair and single-cell resolution. The analysis of large datasets obtained from scBS requires preprocessing to reduce the data size, improve the signal-to-noise ratio and provide interpretability. Typically, this is achieved by dividing the genome into large tiles and averaging the methylation signals within each tile. Here we demonstrate that this coarse-graining approach can lead to signal dilution. We propose improved strategies to identify more informative regions for methylation quantification and a more accurate quantitation method than simple averaging. Our approach enables better discrimination of cell types and other features of interest and reduces the need for large numbers of cells. We also present an approach to detect differentially methylated regions between groups of cells and demonstrate its ability to identify biologically meaningful regions that are associated with genes involved in the core functions of specific cell types. Finally, we present the software tool MethSCAn for scBS data analysis ( https://anders-biostat.github.io/MethSCAn ).
RESUMO
Gene-expression programs define shared and species-specific phenotypes, but their evolution remains largely uncharacterized beyond the transcriptome layer1. Here we report an analysis of the co-evolution of translatomes and transcriptomes using ribosome-profiling and matched RNA-sequencing data for three organs (brain, liver and testis) in five mammals (human, macaque, mouse, opossum and platypus) and a bird (chicken). Our within-species analyses reveal that translational regulation is widespread in the different organs, in particular across the spermatogenic cell types of the testis. The between-species divergence in gene expression is around 20% lower at the translatome layer than at the transcriptome layer owing to extensive buffering between the expression layers, which especially preserved old, essential and housekeeping genes. Translational upregulation specifically counterbalanced global dosage reductions during the evolution of sex chromosomes and the effects of meiotic sex-chromosome inactivation during spermatogenesis. Despite the overall prevalence of buffering, some genes evolved faster at the translatome layer-potentially indicating adaptive changes in expression; testis tissue shows the highest fraction of such genes. Further analyses incorporating mass spectrometry proteomics data establish that the co-evolution of transcriptomes and translatomes is reflected at the proteome layer. Together, our work uncovers co-evolutionary patterns and associated selective forces across the expression layers, and provides a resource for understanding their interplay in mammalian organs.
Assuntos
Evolução Molecular , Mamíferos/genética , Biossíntese de Proteínas , Transcriptoma/genética , Animais , Encéfalo/metabolismo , Galinhas/genética , Feminino , Genes Ligados ao Cromossomo X/genética , Humanos , Fígado/metabolismo , Macaca/genética , Masculino , Camundongos , Gambás/genética , Especificidade de Órgãos/genética , Ornitorrinco/genética , Biossíntese de Proteínas/genética , RNA-Seq , Ribossomos/metabolismo , Cromossomos Sexuais/genética , Especificidade da Espécie , Espermatogênese/genética , Testículo/metabolismo , Regulação para CimaRESUMO
Throughout the SARS-CoV-2 pandemic, limited diagnostic capacities prevented sentinel testing, demonstrating the need for novel testing infrastructures. Here, we describe the setup of a cost-effective platform that can be employed in a high-throughput manner, which allows surveillance testing as an acute pandemic control and preparedness tool, exemplified by SARS-CoV-2 diagnostics in an academic environment. The strategy involves self-sampling based on gargling saline, pseudonymized sample handling, automated RNA extraction, and viral RNA detection using a semiquantitative multiplexed colorimetric reverse transcription loop-mediated isothermal amplification (RT-LAMP) assay with an analytical sensitivity comparable with RT-qPCR. We provide standard operating procedures and an integrated software solution for all workflows, including sample logistics, analysis by colorimetry or sequencing, and communication of results. We evaluated factors affecting the viral load and the stability of gargling samples as well as the diagnostic sensitivity of the RT-LAMP assay. In parallel, we estimated the economic costs of setting up and running the test station. We performed > 35,000 tests, with an average turnover time of < 6 h from sample arrival to result announcement. Altogether, our work provides a blueprint for fast, sensitive, scalable, cost- and labor-efficient RT-LAMP diagnostics, which is independent of potentially limiting clinical diagnostics supply chains.
Assuntos
COVID-19 , SARS-CoV-2 , Humanos , SARS-CoV-2/genética , COVID-19/diagnóstico , COVID-19/epidemiologia , Teste para COVID-19 , Técnicas de Laboratório Clínico/métodos , Pandemias/prevenção & controle , Sensibilidade e Especificidade , RNA Viral/genéticaRESUMO
The evolution of gene expression in mammalian organ development remains largely uncharacterized. Here we report the transcriptomes of seven organs (cerebrum, cerebellum, heart, kidney, liver, ovary and testis) across developmental time points from early organogenesis to adulthood for human, rhesus macaque, mouse, rat, rabbit, opossum and chicken. Comparisons of gene expression patterns identified correspondences of developmental stages across species, and differences in the timing of key events during the development of the gonads. We found that the breadth of gene expression and the extent of purifying selection gradually decrease during development, whereas the amount of positive selection and expression of new genes increase. We identified differences in the temporal trajectories of expression of individual genes across species, with brain tissues showing the smallest percentage of trajectory changes, and the liver and testis showing the largest. Our work provides a resource of developmental transcriptomes of seven organs across seven species, and comparative analyses that characterize the development and evolution of mammalian organs.
Assuntos
Regulação da Expressão Gênica no Desenvolvimento , Organogênese/genética , Transcriptoma/genética , Animais , Evolução Biológica , Galinhas/genética , Feminino , Humanos , Macaca mulatta/genética , Masculino , Camundongos , Gambás/genética , Coelhos , RatosRESUMO
Dimension-reduction methods, such as t-SNE or UMAP, are widely used when exploring high-dimensional data describing many entities, for example, RNA-seq data for many single cells. However, dimension reduction is commonly prone to introducing artifacts, and we hence need means to see where a dimension-reduced embedding is a faithful representation of the local neighborhood and where it is not. We present Sleepwalk, a simple but powerful tool that allows the user to interactively explore an embedding, using color to depict original or any other distances from all points to the cell under the mouse cursor. We show how this approach not only highlights distortions but also reveals otherwise hidden characteristics of the data, and how Sleepwalk's comparative modes help integrate multisample data and understand differences between embedding and preprocessing methods. Sleepwalk is a versatile and intuitive tool that unlocks the full power of dimension reduction and will be of value not only in single-cell RNA-seq but also in any other area with matrix-shaped big data.
Assuntos
RNA-Seq/métodos , Software , Animais , Cerebelo/embriologia , Cerebelo/metabolismo , Expressão Gênica , Camundongos , Análise de Célula ÚnicaRESUMO
This study concerns large margin nearest neighbors classifier and its multi-metric extension as the efficient approaches for metric learning which aimed to learn an appropriate distance/similarity function for considered case studies. In recent years, many studies in data mining and pattern recognition have demonstrated that a learned metric can significantly improve the performance in classification, clustering and retrieval tasks. The paper describes application of the metric learning approach to in silico assessment of chemical liabilities. Chemical liabilities, such as adverse effects and toxicity, play a significant role in drug discovery process, in silico assessment of chemical liabilities is an important step aimed to reduce costs and animal testing by complementing or replacing in vitro and in vivo experiments. Here, to our knowledge for the first time, a distance-based metric learning procedures have been applied for in silico assessment of chemical liabilities, the impact of metric learning on structure-activity landscapes and predictive performance of developed models has been analyzed, the learned metric was used in support vector machines. The metric learning results have been illustrated using linear and non-linear data visualization techniques in order to indicate how the change of metrics affected nearest neighbors relations and descriptor space.
Assuntos
Inteligência Artificial , Simulação por Computador , Relação Estrutura-Atividade , Testes de Carcinogenicidade , Análise por Conglomerados , Canal de Potássio ERG1 , Canais de Potássio Éter-A-Go-Go/antagonistas & inibidores , Inibidores do Fator Xa , Modelos Teóricos , Análise de Componente Principal , SoftwareRESUMO
In research involving data-rich assays, exploratory data analysis is a crucial step. Typically, this involves jumping back and forth between visualizations that provide overview of the whole data and others that dive into details. For example, it might be helpful to have one chart showing a summary statistic for all samples, while a second chart provides details for points selected in the first chart. We present R/LinkedCharts, a framework that renders this task radically simple, requiring very few lines of code to obtain complex and general visualization, which later can be polished to provide interactive data access of publication quality.
Assuntos
Análise de Dados , Software , Interpretação Estatística de Dados , BioensaioRESUMO
In metastasis, cancer cells travel around the circulation to colonize distant sites. Due to the rarity of these events, the immediate fates of metastasizing tumor cells (mTCs) are poorly understood while the role of the endothelium as a dissemination interface remains elusive. Using a newly developed combinatorial mTC enrichment approach, we provide a transcriptional blueprint of the early colonization process. Following their arrest at the metastatic site, mTCs were found to either proliferate intravascularly or extravasate, thereby establishing metastatic latency. Endothelial-derived angiocrine Wnt factors drive this bifurcation, instructing mTCs to follow the extravasation-latency route. Surprisingly, mTC responsiveness towards niche-derived Wnt was established at the epigenetic level, which predetermined tumor cell behavior. Whereas hypomethylation enabled high Wnt activity leading to metastatic latency, methylated mTCs exhibited low activity and proliferated intravascularly. Collectively the data identify the predetermined methylation status of disseminated tumor cells as a key regulator of mTC behavior in the metastatic niche.
Assuntos
Neoplasias Pulmonares , Humanos , Neoplasias Pulmonares/secundário , Neoplasias Pulmonares/patologia , Neoplasias Pulmonares/metabolismo , Animais , Metilação de DNA , Metástase Neoplásica , Camundongos , Linhagem Celular Tumoral , Pulmão/patologia , Proliferação de Células , Proteínas Wnt/metabolismo , Epigênese Genética , Células Neoplásicas Circulantes/patologia , Células Neoplásicas Circulantes/metabolismo , Regulação Neoplásica da Expressão GênicaRESUMO
[This corrects the article DOI: 10.2196/44204.].
RESUMO
Sexually dimorphic traits are common among mammals and are specified during development through the deployment of sex-specific genetic programs. Because little is known about these programs, we investigated them using a resource of gene expression profiles in males and females throughout the development of five organs in five mammals (human, mouse, rat, rabbit, and opossum) and a bird (chicken). We found that sex-biased gene expression varied considerably across organs and species and was often cell-type specific. Sex differences increased abruptly around sexual maturity instead of increasing gradually during organ development. Finally, sex-biased gene expression evolved rapidly at the gene level, with differences between organs in the evolutionary mechanisms used, but more slowly at the cellular level, with the same cell types being sexually dimorphic across species.
Assuntos
Evolução Molecular , Regulação da Expressão Gênica no Desenvolvimento , Mamíferos , Organogênese , Caracteres Sexuais , Animais , Feminino , Humanos , Masculino , Camundongos , Coelhos , Ratos , Galinhas , Mamíferos/genética , Mamíferos/crescimento & desenvolvimento , RNA-Seq , Transcriptoma , Organogênese/genéticaRESUMO
BACKGROUND: The COVID-19 pandemic is characterized by rapid increases in infection burden owing to the emergence of new variants with higher transmissibility and immune escape. To date, monitoring the COVID-19 pandemic has mainly relied on passive surveillance, yielding biased epidemiological measures owing to the disproportionate number of undetected asymptomatic cases. Active surveillance could provide accurate estimates of the true prevalence to forecast the evolution of the pandemic, enabling evidence-based decision-making. OBJECTIVE: This study compared 4 different approaches of active SARS-CoV-2 surveillance focusing on feasibility and epidemiological outcomes. METHODS: A 2-factor factorial randomized controlled trial was conducted in 2020 in a German district with 700,000 inhabitants. The epidemiological outcome comprised SARS-CoV-2 prevalence and its precision. The 4 study arms combined 2 factors: individuals versus households and direct testing versus testing conditioned on symptom prescreening. Individuals aged ≥7 years were eligible. Altogether, 27,908 addresses from 51 municipalities were randomly allocated to the arms and 15 consecutive recruitment weekdays. Data collection and logistics were highly digitized, and a website in 5 languages enabled low-barrier registration and tracking of results. Gargle sample collection kits were sent by post. Participants collected a gargle sample at home and mailed it to the laboratory. Samples were analyzed with reverse transcription loop-mediated isothermal amplification (RT-LAMP); positive and weak results were confirmed with real-time reverse transcription-polymerase chain reaction (RT-PCR). RESULTS: Recruitment was conducted between November 18 and December 11, 2020. The response rates in the 4 arms varied between 34.31% (2340/6821) and 41.17% (2043/4962). The prescreening classified 16.61% (1207/7266) of the patients as COVID-19 symptomatic. Altogether, 4232 persons without prescreening and 7623 participating in the prescreening provided 5351 gargle samples, of which 5319 (99.4%) could be analyzed. This yielded 17 confirmed SARS-CoV-2 infections and a combined prevalence of 0.36% (95% CI 0.14%-0.59%) in the arms without prescreening and 0.05% (95% CI 0.00%-0.108%) in the arms with prescreening (initial contacts only). Specifically, we found a prevalence of 0.31% (95% CI 0.06%-0.58%) for individuals and 0.35% (95% CI 0.09%-0.61%) for households, and lower estimates with prescreening (0.07%, 95% CI 0.0%-0.15% for individuals and 0.02%, 95% CI 0.0%-0.06% for households). Asymptomatic infections occurred in 27% (3/11) of the positive cases with symptom data. The 2 arms without prescreening performed the best regarding effectiveness and accuracy. CONCLUSIONS: This study showed that postal mailing of gargle sample kits and returning home-based self-collected liquid gargle samples followed by high-sensitivity RT-LAMP analysis is a feasible way to conduct active SARS-CoV-2 population surveillance without burdening routine diagnostic testing. Efforts to improve participation rates and integration into the public health system may increase the potential to monitor the course of the pandemic. TRIAL REGISTRATION: Deutsches Register Klinischer Studien (DRKS) DRKS00023271; https://tinyurl.com/3xenz68a. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): RR2-10.1186/s13063-021-05619-5.
Assuntos
COVID-19 , SARS-CoV-2 , Humanos , COVID-19/epidemiologia , Pandemias/prevenção & controle , Manejo de Espécimes , LaboratóriosRESUMO
BACKGROUND: To achieve higher effectiveness in population-based SARS-CoV-2 surveillance and to reliably predict the course of an outbreak, screening, and monitoring of infected individuals without major symptoms (about 40% of the population) will be necessary. While current testing capacities are also used to identify such asymptomatic cases, this rather passive approach is not suitable in generating reliable population-based estimates of the prevalence of asymptomatic carriers to allow any dependable predictions on the course of the pandemic. METHODS: This trial implements a two-factorial, randomized, controlled, multi-arm, prospective, interventional, single-blinded design with cluster sampling and four study arms, each representing a different SARS-CoV-2 testing and surveillance strategy based on individuals' self-collection of saliva samples which are then sent to and analyzed by a laboratory. The targeted sample size for the trial is 10,000 saliva samples equally allocated to the four study arms (2500 participants per arm). Strategies differ with respect to tested population groups (individuals vs. all household members) and testing approach (without vs. with pre-screening survey). The trial is complemented by an economic evaluation and qualitative assessment of user experiences. Primary outcomes include costs per completely screened person, costs per positive case, positive detection rate, and precision of positive detection rate. DISCUSSION: Systems for active surveillance of the general population will gain more importance in the context of pandemics and related disease prevention efforts. The pandemic parameters derived from such active surveillance with routine population monitoring therefore not only enable a prospective assessment of the short-term course of a pandemic, but also a more targeted and thus more effective use of local and short-term countermeasures. TRIAL REGISTRATION: ClinicalTrials.gov DRKS00023271 . Registered November 30, 2020, with the German Clinical Trials Register (Deutsches Register Klinischer Studien).
Assuntos
COVID-19 , SARS-CoV-2 , Teste para COVID-19 , Análise Custo-Benefício , Humanos , Grupos Populacionais , Estudos Prospectivos , Ensaios Clínicos Controlados Aleatórios como Assunto , Resultado do TratamentoRESUMO
OBJECTIVES: In this cluster-randomised controlled study (CoV-Surv Study), four different "active" SARS-CoV-2 testing strategies for general population surveillance are evaluated for their effectiveness in determining and predicting the prevalence of SARS-CoV-2 infections in a given population. In addition, the costs and cost-effectiveness of the four surveillance strategies will be assessed. Further, this trial is supplemented by a qualitative component to determine the acceptability of each strategy. Findings will inform the choice of the most effective, acceptable and affordable strategy for SARS-CoV-2 surveillance, with the most effective and cost-effective strategy becoming part of the local public health department's current routine health surveillance activities. Investigating its everyday performance will allow us to examine the strategy's applicability to real time prevalence prediction and the usefulness of the resulting information for local policy makers to implement countermeasures that effectively prevent future nationwide lockdowns. The authors would like to emphasize the importance and relevance of this study and its expected findings in the context of population-based disease surveillance, especially in respect to the current SARS-CoV-2 pandemic. In Germany, but also in many other countries, COVID-19 surveillance has so far largely relied on passive surveillance strategies that identify individuals with clinical symptoms, monitor those cases who then tested positive for the virus, followed by tracing of individuals in close contact to those positive cases. To achieve higher effectiveness in population surveillance and to reliably predict the course of an outbreak, screening and monitoring of infected individuals without major symptoms (about 40% of the population) will be necessary. While current testing capacities are also used to identify such asymptomatic cases, this rather passive approach is not suitable in generating reliable population-based estimates of the prevalence of asymptomatic carriers to allow any dependable predictions on the course of the pandemic. To better control and manage the SARS-CoV-2 pandemic, current strategies therefore need to be complemented by an active surveillance of the wider population, i.e. routinely conducted testing and monitoring activities to identify and isolate infected individuals regardless of their clinical symptoms. Such active surveillance strategies will enable more effective prevention of the spread of the virus as they can generate more precise population-based parameters during a pandemic. This essential information will be required in order to determine the best strategic and targeted short-term countermeasures to limit infection spread locally. TRIAL DESIGN: This trial implements a cluster-randomised, two-factorial controlled, prospective, interventional, single-blinded design with four study arms, each representing a different SARS-CoV-2 testing and surveillance strategy. PARTICIPANTS: Eligible are individuals age 7 years or older living in Germany's Rhein-Neckar Region who consent to provide a saliva sample (all four arms) after completion of a brief questionnaire (two arms only). For the qualitative component, different samples of study participants and non-participants (i.e. eligible for study, but refuse to participate) will be identified for additional interviews. For these interviews, only individuals age 18 years or older are eligible. INTERVENTION AND COMPARATOR: Of the four surveillance strategies to be assessed and compared, Strategy A1 is considered the gold standard for prevalence estimation and used to determine bias in other arms. To determine the cost-effectiveness, each strategy is compared to status quo, defined as the currently practiced passive surveillance approach. Strategy A1: Individuals (one per household) receive information and study material by mail with instructions on how to produce a saliva sample and how to return the sample by mail. Once received by the laboratory, the sample is tested for SARS-CoV-2 using Reverse Transcription Loop-mediated Isothermal Amplification (RT-LAMP). Strategy A2: Individuals (one per household) receive information and study material by mail with instructions on how to produce their own as well as saliva samples from each household member and how to return these samples by mail. Once received by the laboratory, the samples are tested for SARS-CoV-2 using RT-LAMP. Strategy B1: Individuals (one per household) receive information by mail on how to complete a brief pre-screening questionnaire which asks about COVID-19 related clinical symptoms and risk exposures. Only individuals whose pre-screening score crosses a defined threshold, will then receive additional study material by mail with instructions on how to produce a saliva sample and how to return the sample by mail. Once received by the laboratory, the saliva sample is tested for SARS-CoV-2 using RT-LAMP. Strategy B2: Individuals (one per household) receive information by mail on how to complete a brief pre-screening questionnaire which asks about COVID-19 related clinical symptoms. Only individuals whose pre-screening score crosses a defined threshold, will then receive additional study material by mail with instructions how to produce their own as well as saliva samples from each household member and how to return these samples by mail. Once received by the laboratory, the samples are tested for SARS-CoV-2 using RT-LAMP. In each strategy, RT-LAMP positive samples are additionally analyzed with qPCR in order to minimize the number of false positives. MAIN OUTCOMES: The identification of the one best strategy will be determined by a set of parameters. Primary outcomes include costs per correctly screened person, costs per positive case, positive detection rate, and precision of positive detection rate. Secondary outcomes include participation rate, costs per asymptomatic case, prevalence estimates, number of asymptomatic cases per study arm, ratio of symptomatic to asymptomatic cases per study arm, participant satisfaction. Additional study components (not part of the trial) include cost effectiveness of each of the four surveillance strategies compared to passive monitoring (i.e. status quo), development of a prognostic model to predict hospital utilization caused by SARS-CoV-2, time from test shipment to test application and time from test shipment to test result, and perception and preferences of the persons to be tested with regard to test strategies. RANDOMISATION: Samples are drawn in three batches of three continuous weeks. Randomisation follows a two-stage process. First, a total of 220 sampling points have been allocated to the three different batches. To obtain an integer solution, the Cox-algorithm for controlled rounding has been used. Afterwards, sample points have been drawn separately per batch, following a probability proportional to size (PPS) random sample. Second, for each cluster the same number of residential addresses is randomly sampled from the municipal registries (self-weighted sample of individuals). The 28,125 addresses drawn per municipality are then randomly allocated to the four study arms A1, A2, B1, and B2 in the ratio 5 to 2.5 to 14 to 7 based on the expected response rates in each arm and the sensitivity and specificity of the pre-screening tool as applied in strategy B1 and B2. Based on the assumptions, this allocation should yield 2500 saliva samples in each strategy. Although a municipality can be sampled by multiple batches and the overall number of addresses per municipality might vary, the number of addresses contacted in each arm is kept constant. BLINDING (MASKING): The design is single-blinded, meaning the staff conducting the SARS-CoV-2 tests are unaware of the study arm assignment of each single participant and test sample. SAMPLE SIZES: Total sample size for the trial is 10,000 saliva samples equally allocated to the four study arms (i.e. 2,500 participants per arm). For the qualitative component, up to 60 in-depth interviews will be conducted with about 30 study participants (up to 15 in each arm A and B) and 30 participation refusers (up to 15 in each arm A and B) purposefully selected from the quantitative study sample to represent a variety of gender and ages to explore experiences with admission or rejection of study participation. Up to 25 asymptomatic SARS-CoV-2 positive study participants will be purposefully selected to explore the way in which asymptomatic men and women diagnosed with SARS-CoV-2 give meaning to their diagnosis and to the dialectic between feeling concurrently healthy and yet also being at risk for transmitting COVID-19. In addition, 100 randomly selected study participants will be included to explore participants' perspective on testing processes and implementation. TRIAL STATUS: Final protocol version is "Surveillance_Studienprotokoll_03Nov2020_v1_2" from November 3, 2020. Recruitment started November 18, 2020 and is expected to end by or before December 31, 2020. TRIAL REGISTRATION: The trial is currently being registered with the German Clinical Trials Register (Deutsches Register Klinischer Studien), DRKS00023271 ( https://www.drks.de/drks_web/navigate.do?navigationId=trial . HTML&TRIAL_ID=DRKS00023271). Retrospectively registered 30 November 2020. FULL PROTOCOL: The full protocol is attached as an additional file, accessible from the Trials website (Additional file 1). In the interest in expediting dissemination of this material, the familiar formatting has been eliminated; this Letter serves as a summary of the key elements of the full protocol.
Assuntos
Teste de Ácido Nucleico para COVID-19/economia , COVID-19/diagnóstico , COVID-19/economia , Custos de Cuidados de Saúde , Técnicas de Diagnóstico Molecular/economia , Técnicas de Amplificação de Ácido Nucleico/economia , SARS-CoV-2/genética , Saliva/virologia , Inquéritos e Questionários/economia , COVID-19/epidemiologia , COVID-19/virologia , Análise Custo-Benefício , Feminino , Alemanha/epidemiologia , Humanos , Masculino , Vigilância da População , Valor Preditivo dos Testes , Prevalência , Ensaios Clínicos Controlados Aleatórios como Assunto , Reprodutibilidade dos Testes , Método Simples-CegoRESUMO
An amendment to this paper has been published and can be accessed via a link at the top of the paper.
Assuntos
Betacoronavirus/patogenicidade , Infecções por Coronavirus/epidemiologia , Pandemias/estatística & dados numéricos , Pneumonia Viral/epidemiologia , Inquéritos e Questionários/estatística & dados numéricos , COVID-19 , Infecções por Coronavirus/prevenção & controle , Infecções por Coronavirus/virologia , Nível de Saúde , Humanos , Pandemias/prevenção & controle , Pneumonia Viral/prevenção & controle , Pneumonia Viral/virologia , SARS-CoV-2RESUMO
Chemical liabilities, such as adverse effects and toxicity, play a significant role in modern drug discovery process. In silico assessment of chemical liabilities is an important step aimed to reduce costs and animal testing by complementing or replacing in vitro and in vivo experiments. Herein, we propose an approach combining several classification and chemography methods to be able to predict chemical liabilities and to interpret obtained results in the context of impact of structural changes of compounds on their pharmacological profile. To our knowledge for the first time, the supervised extension of Generative Topographic Mapping is proposed as an effective new chemography method. New approach for mapping new data using supervised Isomap without re-building models from the scratch has been proposed. Two approaches for estimation of model's applicability domain are used in our study to our knowledge for the first time in chemoinformatics. The structural alerts responsible for the negative characteristics of pharmacological profile of chemical compounds has been found as a result of model interpretation.
RESUMO
Over the years, a number of dimensionality reduction techniques have been proposed and used in chemoinformatics to perform nonlinear mappings. In this study, four representatives of nonlinear dimensionality reduction methods related to two different families were analyzed: distance-based approaches (Isomap and Diffusion Maps) and topology-based approaches (Generative Topographic Mapping (GTM) and Laplacian Eigenmaps). The considered methods were applied for the visualization of three toxicity datasets by using four sets of descriptors. Two methods, GTM and Diffusion Maps, were identified as the best approaches, which thus made it impossible to prioritize a single family of the considered dimensionality reduction methods. The intrinsic dimensionality assessment of data was performed by using the Maximum Likelihood Estimation. It was observed that descriptor sets with a higher intrinsic dimensionality contributed maps of lower quality. A new statistical coefficient, which combines two previously known ones, was proposed to automatically rank the maps. Instead of relying on one of the best methods, we propose to automatically generate maps with different parameter values for different descriptor sets. By following this procedure, the maps with the highest values of the introduced statistical coefficient can be automatically selected and used as a starting point for visual inspection by the user.